@inproceedings{lyu-etal-2020-revisiting,
title = "Revisiting {M}odularized {M}ultilingual {NMT} to {M}eet {I}ndustrial {D}emands",
author = "Lyu, Sungwon and
Son, Bokyung and
Yang, Kichang and
Bae, Jaekyoung",
editor = "Webber, Bonnie and
Cohn, Trevor and
He, Yulan and
Liu, Yang",
booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)",
month = nov,
year = "2020",
address = "Online",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2020.emnlp-main.476/",
doi = "10.18653/v1/2020.emnlp-main.476",
pages = "5905--5918",
abstract = "The complete sharing of parameters for multilingual translation (1-1) has been the mainstream approach in current research. However, degraded performance due to the capacity bottleneck and low maintainability hinders its extensive adoption in industries. In this study, we revisit the multilingual neural machine translation model that only share modules among the same languages (M2) as a practical alternative to 1-1 to satisfy industrial requirements. Through comprehensive experiments, we identify the benefits of multi-way training and demonstrate that the M2 can enjoy these benefits without suffering from the capacity bottleneck. Furthermore, the interlingual space of the M2 allows convenient modification of the model. By leveraging trained modules, we find that incrementally added modules exhibit better performance than singly trained models. The zero-shot performance of the added modules is even comparable to supervised models. Our findings suggest that the M2 can be a competent candidate for multilingual translation in industries."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="lyu-etal-2020-revisiting">
<titleInfo>
<title>Revisiting Modularized Multilingual NMT to Meet Industrial Demands</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sungwon</namePart>
<namePart type="family">Lyu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bokyung</namePart>
<namePart type="family">Son</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kichang</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jaekyoung</namePart>
<namePart type="family">Bae</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2020-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Bonnie</namePart>
<namePart type="family">Webber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Trevor</namePart>
<namePart type="family">Cohn</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yulan</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Online</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>The complete sharing of parameters for multilingual translation (1-1) has been the mainstream approach in current research. However, degraded performance due to the capacity bottleneck and low maintainability hinders its extensive adoption in industries. In this study, we revisit the multilingual neural machine translation model that only share modules among the same languages (M2) as a practical alternative to 1-1 to satisfy industrial requirements. Through comprehensive experiments, we identify the benefits of multi-way training and demonstrate that the M2 can enjoy these benefits without suffering from the capacity bottleneck. Furthermore, the interlingual space of the M2 allows convenient modification of the model. By leveraging trained modules, we find that incrementally added modules exhibit better performance than singly trained models. The zero-shot performance of the added modules is even comparable to supervised models. Our findings suggest that the M2 can be a competent candidate for multilingual translation in industries.</abstract>
<identifier type="citekey">lyu-etal-2020-revisiting</identifier>
<identifier type="doi">10.18653/v1/2020.emnlp-main.476</identifier>
<location>
<url>https://aclanthology.org/2020.emnlp-main.476/</url>
</location>
<part>
<date>2020-11</date>
<extent unit="page">
<start>5905</start>
<end>5918</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Revisiting Modularized Multilingual NMT to Meet Industrial Demands
%A Lyu, Sungwon
%A Son, Bokyung
%A Yang, Kichang
%A Bae, Jaekyoung
%Y Webber, Bonnie
%Y Cohn, Trevor
%Y He, Yulan
%Y Liu, Yang
%S Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
%D 2020
%8 November
%I Association for Computational Linguistics
%C Online
%F lyu-etal-2020-revisiting
%X The complete sharing of parameters for multilingual translation (1-1) has been the mainstream approach in current research. However, degraded performance due to the capacity bottleneck and low maintainability hinders its extensive adoption in industries. In this study, we revisit the multilingual neural machine translation model that only share modules among the same languages (M2) as a practical alternative to 1-1 to satisfy industrial requirements. Through comprehensive experiments, we identify the benefits of multi-way training and demonstrate that the M2 can enjoy these benefits without suffering from the capacity bottleneck. Furthermore, the interlingual space of the M2 allows convenient modification of the model. By leveraging trained modules, we find that incrementally added modules exhibit better performance than singly trained models. The zero-shot performance of the added modules is even comparable to supervised models. Our findings suggest that the M2 can be a competent candidate for multilingual translation in industries.
%R 10.18653/v1/2020.emnlp-main.476
%U https://aclanthology.org/2020.emnlp-main.476/
%U https://doi.org/10.18653/v1/2020.emnlp-main.476
%P 5905-5918
Markdown (Informal)
[Revisiting Modularized Multilingual NMT to Meet Industrial Demands](https://aclanthology.org/2020.emnlp-main.476/) (Lyu et al., EMNLP 2020)
ACL